How to (really) read?

[insaneinthemembrane]

Hello everyone

 

I'm a first year graduate student in the biosciences coming directly from undergraduate. I'm struggling with learning how to really read the amount of papers that are required weekly for class on top of reading for my rotation lab to develop an NSF proposal. I was wondering if any senior grad students or postdocs had any advice on how to read through large volumes of papers while still coming away with lots of knowledge. Currently, it take me anywhere between an hour to 1 1/2 hours to fully read and annotate a paper. Any tips on how to do it faster?

[rising_star]

The quick answer is not to fully read and annotate everything. Read for what you need, rather than just reading in general. For example, if you're trying to understand different methods used to study a topic, then you'll want to focus and take notes on the methods sections. If you're concerned with results, then focus on that, making sure to pay attention to any figures and tables included.

Also, stop trying to annotate everything. You're better off writing a brief (1-2 paragraphs, 300-400 word max) summary of the article and its key points in your own words, so you can easily do a search to find the relevant stuff later. No one really reads every single article carefully from first page to last page. 

[dr. t]

 

Edited October 8, 2015 by telkanuru

[TakeruK]

I would strongly second rising_star's advice against annotating everything. This is how I started reading in my first year of grad school too and it didn't help me at all. Summarizing in your own words and maybe even presenting a 5 minute summary of it to a colleague is what really helps me understand material.

There are still maybe a handful (less than 5 papers) that I've read every single word, annotated the margins, and can recite full equations from. These papers are the ones that form the foundation of the methods and ideas behind my main research questions so I think it's important to know that at this level. But you should not be doing this for every paper. Also, for the papers that I did annotate everything, I did the quick thing rising_star and telkanuru's PhD comic said first, I just "went through" them and after going through a lot of papers, I was able to identify the key foundational works and then read them more deeply.

[insaneinthemembrane]

Thanks for the advice everyone! I will try and incorporate it into my readings now. 

[Char Tutor:Mentor]

The quick answer is not to fully read and annotate everything. Read for what you need, rather than just reading in general. For example, if you're trying to understand different methods used to study a topic, then you'll want to focus and take notes on the methods sections. If you're concerned with results, then focus on that, making sure to pay attention to any figures and tables included.

Also, stop trying to annotate everything. You're better off writing a brief (1-2 paragraphs, 300-400 word max) summary of the article and its key points in your own words, so you can easily do a search to find the relevant stuff later. No one really reads every single article carefully from first page to last page. 

[GGeorgian]

First of all, learn how to skim the text, it will help you a lot, trust me. 

Annotation for everything will make you tired after even ten page readings, so try to make the comments alongside the text ONLY on whatever really matters to the subject you are trying to learn. 

And one more practice that I find extremely useful, is the following: from the initial stage of reading/skimming imagine you are going to do a presentation on that topic/subject. UNDERLINE all those words/sentences/listings whatever you think will be appropriate to be situated on the presentation slides. Keep in mind that presentations are never loaded with lots of texts. It should be maximum of 12 words in a sentence, but always try to make it smaller. In a matter of comments alongside the text, you can write the words or even small sentences regarding the topic, whatever you virtually think that will include in "presentation"

Once you master this technique, you'll naturally start to annotate/underline only the necessary stuff, what can keep on track to learn/revise the topic.

[zipykido]

I have a pretty simple checklist for reading papers for assignments. This changes a bit if I'm reading a paper for review or for my own research but the steps are pretty similar. 

1. Read abstract. You'll get most of the information you'll need out of here if the paper is well written.

2. Read introduction. If the topic is unfamiliar to you then make sure you understand the introduction which provides the framework for the entire paper usually includes motivation, past work, and gaps in scientific knowledge.

3. Read discussion/conclusion. Note significant claims made by the author, this will be important for the following steps.

4. Figure out the figures. This is the only step in which you can only get better with practice. Once you start recognizing data presentation, reading goes a lot faster. For instance I can skim flow charts and survival charts very quickly now without having to refer to staining antibodies or conditions. But when I started it took forever to do it, and would slow down my reading to a point where it was frustrating. Make sure that the data is convincing and supports the claims made in the discussion (statistics and presentation matter!). Conclusions should be drawn logically from the data and not randomly guessed or buried in graphs and tables. Ask yourself if you were presented with this data, would you have the same conclusion? Well laid out papers will have logical sense, figure 1 will present something, figure 2 will fill in holes from figure 1 and so on. 

5. Skim text. Methods and written text are important but only so much as to understand the data presented. If the figures are confusing and not well made then you'll have to dig into the text a bit to figure out what they did and hopefully why they did it. Once you're more ingrained in your field you'll be able to spot subtleties in methods that will skew the data in one way or the other. This is important for point 4.

Also note that not all papers are well written or have significant contribution to the field. Don't get bogged down by those papers. I've read papers with 8 figures while only 2 of them were necessary. Once you understand the submission process and the politics behind papers you'll realize that some papers that have been accepted are just terrible. Just read to come out with key findings and key figures to support or refute your argument. 

[mikeariehmed]

I hope this article will help. http://gradschooljungle.blogspot.com/2013/01/the-truths-of-reading.html

[DawnGleamTurtle]

I've had trouble reading academic literature since my undergrad years, and so I gotta say, thank you for this thread.

I've been a grad student long enough to have eventually stumbled on the truth that reading every word and treating the text as some sacred fount of knowledge is a load of crap.  Furthermore, reading stuff written about a very specific topic tends to cause every intro section to say roughly the same things.

On 11/3/2015, 4:16:22, mikeariehmed said:

I hope this article will help. http://gradschooljungle.blogspot.com/2013/01/the-truths-of-reading.html

This article is definitely helpful.  Though I'm not sure how good that last one is -- for me I think I can't really parse text without hearing the word in my mind.  If anything, what actually happens is that if I need to read fast I just dart from one part of the paragraph to the next to get a general idea of what's going on from some semi-random selection of words that look important.  (Okay, that's not random at all, I admit.)

I do sometimes suffer "reader's block" where I sit in front of a paper or two or an entire stack and just keep getting distracted.  I can't seem to find any better way to deal with it other than to simply suck it up and read the paper.  And I try to tell myself if I've read even a tiny fraction of the content in the paper, it's still better than having read none of it.  To some extent, the literature feels a little like a Mega Man game, in that there's a familiarity with the material that would allow me to read these papers smoothly and easily, but I can't get that familiarity without first breaking into this bunch of papers, just like how all the robot masters in Mega Man games have weapons that other robot masters are weak to, but you've just got to get your first robot master weapon with your own default weapon.

If anyone's got a better idea than "just read it, as much or as little as you can", please do let me know.

On 10/8/2015, 10:01:52, telkanuru said:

Hahah this is hilarious.  And quite probably true.  Which makes me wonder why people like to be so wordy and put so much into prose.

Would be nice if papers were just a large amount of tables, graphs, flowcharts, bullet-point lists, and such, instead of prose.  Prose is the most annoying thing to extract information from, in my opinion.

Also, long paragraphs.  Ugh.

Problem is that -- if you're great at critical thinking and poking holes in arguments and especially if you're the type to be very thorough and like to cement your understanding before moving on and also like to make sure you know everything about something in case someone tries to poke holes in what you say about it -- you feel this strong impulse that you're missing something and have to go back and read it or else you won't truly understand it.  That feeling -- and the accompanying notion that one has to sit down and get serious with a good chunk of time formally set aside to tackle a major operation known as reading an academic paper, just to be sure that one understands it completely and whole and comprehensively -- is a major contributor to reader's block.

Edited November 18, 2015 by DawnGleamTurtle

[hippyscientist]

I've found that by reading (certain) papers multiple times, it's a lot easier to pick holes in them/critically evaluate the arguments they put forward. I'm not suggesting to do that for every bit of reading you need to do, but for certain assignments/key literature it's quite useful.

I get on average 25 papers a week that I have to read for my classes, and then more that I end up reading so my assignments are good and my research makes sense. I would say I'm reading approximately 40 papers a week right now, and so I've learnt to find the relevant parts. If I'm reading for background knowledge, introductions and discussions are key. If I'm reading to understand methodologies, or how to present my own more succinctly, I read the methods. If I'm unsure on the best way to present my results, I just look at the results sections. Sometimes I have to read the whole paper, but by cutting down to the relevant sections you can minimise how many you "read".

 

[ballwera]

Really depends on what I'm reading for, but I've really just started reading abstracts and looking at figures. If you don't understand the figure, read the associated sections. If it's directly relevant for my research I'll read a bit deeper. 

[Concordia]

On 10/8/2015 at 8:48 AM, rising_star said:

 

Also, stop trying to annotate everything. You're better off writing a brief (1-2 paragraphs, 300-400 word max) summary of the article and its key points in your own words, so you can easily do a search to find the relevant stuff later. No one really reads every single article carefully from first page to last page. 

Not to back away from the main topic, but what is the best way to organize those notes so they are searchable and useful?  I'm doing a part-time master's right now, so I haven't been completely crushed by staying on top of the reading, but even I get bothered by a paleolithic kind of note-taking that I'm stuck in.  EasyBib for what I know will be important sources, Word transcripts of the EasyBib file, stored pdfs of articles from JStor on my computer (+ SugarSync backup), and then a chaotic mess of physical notebooks, and various Word documents to capture bits of reading and brilliant ideas of my own I might want to remember.  If I had a better system, I'd be more able to just read without worrying about consequences.

[Neist]

4 minutes ago, Concordia said:

Not to back away from the main topic, but what is the best way to organize those notes so they are searchable and useful?  I'm doing a part-time master's right now, so I haven't been completely crushed by staying on top of the reading, but even I get bothered by a paleolithic kind of note-taking that I'm stuck in.  EasyBib for what I know will be important sources, Word transcripts of the EasyBib file, stored pdfs of articles from JStor on my computer (+ SugarSync backup), and then a chaotic mess of physical notebooks, and various Word documents to capture bits of reading and brilliant ideas of my own I might want to remember.  If I had a better system, I'd be more able to just read without worrying about consequences.

Have you tried PDF annotation software to make one's life easier?

I simply annotate the PDFs themselves. Infinitely easier, once you get the hang of it.

[Concordia]

Is there a particularly useful one I could look up?  And is there a way to further integrate bibliography/citations and search functions?  

Sorry to be so ignorant of this, but I've spent the last few decades in an environment that had a completely different relationship to documents.

 

 

[hippyscientist]

3 minutes ago, Concordia said:

Is there a particularly useful one I could look up?  And is there a way to further integrate bibliography/citations and search functions?  

Sorry to be so ignorant of this, but I've spent the last few decades in an environment that had a completely different relationship to documents.

 

 

My experience of Mendeley is awesome and has all those features you're talking about. Plus it's free

[rising_star]

31 minutes ago, Concordia said:

Not to back away from the main topic, but what is the best way to organize those notes so they are searchable and useful?  I'm doing a part-time master's right now, so I haven't been completely crushed by staying on top of the reading, but even I get bothered by a paleolithic kind of note-taking that I'm stuck in.  EasyBib for what I know will be important sources, Word transcripts of the EasyBib file, stored pdfs of articles from JStor on my computer (+ SugarSync backup), and then a chaotic mess of physical notebooks, and various Word documents to capture bits of reading and brilliant ideas of my own I might want to remember.  If I had a better system, I'd be more able to just read without worrying about consequences.

I type my reading notes/summaries into Zotero, where they're then fully searchable.

[dmsquf12]

Mendeley and NCBI is also good data/research bank for bibliography and citation. Plus, Mendeley is free except some research papers are not available via the software. But, there's always a way that you can get free-access with your school email. 

 

[Neist]

11 minutes ago, Concordia said:

Is there a particularly useful one I could look up?  And is there a way to further integrate bibliography/citations and search functions?  

Sorry to be so ignorant of this, but I've spent the last few decades in an environment that had a completely different relationship to documents.

I can only speak to what I use, but I like Paperpile in combination with Kami. You can import articles into Google Drive straight from your browser (assuming you use Chrome), and Kami allows you to insert citations from your browser directly onto the articles you've imported into Google Drive. All you have to do is import them in with the widget button in Chrome, edit them as you like, then forget about them. Paperpile also sorts the articles by author's name, and you can tag/sort them within the web-based app however you like. Paperpile is also searchable, and it aids in citation insertion and bibliography creation.

I've never had any issues with this strategy yet, and if you use Google products, I'd highly recommend them. Kami has a free version, but Paperpile does cost a monthly fee ($2.99/month for academics). 

Edit: I should note that whatever you use, make sure you can export your work out as either BibTex or RIS. Most reference management software read either of these, so if you want to migrate later on, it should be relatively painless if you can export into these file formats.

Edited March 23, 2016 by Neist

[Concordia]

Thanks, all.  No obvious solution yet, but much to think over.